During their migration, cerebellar granule cells switch from a tangential to a radial mode of migration. We have previously demonstrated that this involves the transmembrane semaphorin Sema6A. We show here that plexin-A2 is the receptor that controls Sema6A function in migrating granule cells. In plexin-A2-deficient (Plxna2(-/-)) mice, which were generated by homologous recombination, many granule cells remained in the molecular layer, as we saw in Sema6a mutants. A similar phenotype was observed in mutant mice that were generated by mutagenesis with N-ethyl-N-nitrosourea and had a single amino-acid substitution in the semaphorin domain of plexin-A2. We found that this mutation abolished the ability of Sema6A to bind to plexin-A2. Mouse chimera studies further suggested that plexin-A2 acts in a cell-autonomous manner. We also provide genetic evidence for a ligand-receptor relationship between Sema6A and plexin-A2 in this system. Using time-lapse video microscopy, we found that centrosome-nucleus coupling and coordinated motility were strongly perturbed in Sema6a(-/-) and Plxna2(-/-) granule cells. This suggests that semaphorin-plexin signaling modulates cell migration by controlling centrosome positioning.

Slits, semaphorins and netrins are three families of proteins that can attract or repel growing axons and migrating neurons in the developing nervous system of vertebrates and invertebrates. Recent studies have shown that they are widely expressed outside the nervous system and that they may play important roles in cancers. Several of the genes encoding these proteins are localized on chromosomal region associated with frequent loss-of-heterozygosity in tumors and cancer cell lines and there is also significant hypermethylation of their promoter suggesting that they may act as tumor suppressors. In addition, proteins in all these families and their receptors appear to control the vascularization of the tumors. Last, many axon guidance molecules also regulate cell migration and apoptosis in normal and tumorigenic tissues. Overall, this suggests that molecules that could mimick or block the activity of axon guidance molecules may be used as therapeutic agents for the treatment of malignancy.

The transmembrane semaphorin protein Sema6A is broadly expressed in the developing nervous system. Sema6A repels several classes of developing axons in vitro and contributes to thalamocortical axon guidance in vivo. Here we show that during cerebellum development, Sema6A is selectively expressed by postmitotic granule cells during their tangential migration in the deep external granule cell layer, but not during their radial migration. In Sema6A-deficient mice, many granule cells remain ectopic in the molecular layer where they differentiate and are contacted by mossy fibers. The analysis of ectopic granule cell morphology in Sema6a-/- mice, and of granule cell migration and neurite outgrowth in cerebellar explants, suggests that Sema6A controls the initiation of granule cell radial migration, probably through a modulation of nuclear and/or soma translocation. Finally, the analysis of mouse chimeras suggests that this function of Sema6A is primarily non-cell-autonomous.